Development of prediction model for axle torque of agricultural tractors

The tractor driving axle torque is an important factor in optimal transmission design and service life evaluation. Axle torque measurement sensor systems are very expensive, and traction force-based axle torque prediction models cannot accurately estimate the axle torque because they do not consider both the conditions of the tractor and the attached implement. Therefore, in this study, a prediction model was developed to estimate the axle torque of an agricultural tractor based on the traction force equation and motion resistance. A load measurement system was established to verify the developed prediction model, and actual field torque data were collected through field tests. The developed prediction model was verified by comparing the results of five reference prediction methods, including weight, engine-rated torque, and three traction equations (Wismer-Luth, ASABE Standard D497.4, and Brixius), using the measured axle torque. Performance evaluation was conducted based on the main variables, including travel speed, tillage depth, and slip ratio. The proposed prediction model was found to be closest to the 1:1 line at all travel speeds, tillage depths, and slip ratios, implying that it can best explain the measured torque values among all prediction models. Compared to the other prediction models, the proposed prediction model's results under all variable conditions had an R2 of 0.65, MAPE of 2.1%, RMSE of 29 Nm, and RD of 2.7%, indicating excellent prediction of the measured torque. The results show that the developed prediction model can be applied to axle torque prediction by explaining the actual measured axle torque.